perf(sky): ~10x cheaper night-sky lattice — one-round cell hash, 2x2x2 block, bit-sliced stars

The owner measured 63% GPU at night vs 28% by day: the starfield
lattice was the whole difference. Three lossless cuts:
- one packed-multiply hash round per cell instead of ~9 chained PCG
  rounds (position/presence bit-sliced from one result, brightness/tint
  from a second);
- the 3x3x3 neighbourhood sweep becomes the 2x2x2 block around the
  sample — a star's visible footprint (a few px) is far smaller than a
  lattice cell on screen, so any star outside that block is over a full
  cell (>=15 px) away and contributes nothing;
- background mottle drops its invisible third octave.

Same densities, sizes, colours, and look; star positions reshuffle
(procedural layout, not authored). sky.frag.spv re-pinned.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>

tests/AcDream.App.Tests/Rendering/Gpu/Vk/VulkanShaderManifestTests.cs

@@ -52,7 +52,7 @@ public sealed class VulkanShaderManifestTests
             // override, retail's GameSky::Draw @0x00506FF0 rule (see
             // SkyFogRuleTests). A deliberate default-path change, reviewed
             // with the world-fog-range fix in the same commit.
-            ["sky.frag.spv"] = "e307a7f3cf84aada86db7c2859c55841a650b46732f12149986562cef0a90ca8",
+            ["sky.frag.spv"] = "1c4ae77056837cbdc188f8cfcc4b0e8851647cdfaf398f25d8c8ff489ef84d57",
             ["sky.vert.spv"] = "3b51945fa4ff1be1604144df92866bdd47aade22f9dd90267591ef36adb28cde",
             ["terrain_modern.frag.spv"] = "7b3cdb01b837ed77ee20559a81c1ce5c9d5395300efcc072560ab0be3c5a1af9",
             ["terrain_modern.vert.spv"] = "9f4cb221ea6aed94a8d23af6cb8e3f3ed96c3cce6e50d135a72d3b55667b1557",